This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Will tree-ssa be GCC 3.5?
- From: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- To: law at redhat dot com
- Cc: gcc at gcc dot gnu dot org
- Date: Mon, 1 Dec 03 23:59:50 EST
- Subject: Re: Will tree-ssa be GCC 3.5?
CSE is an example of a pass that I want to see significantly simplified.
Most of the complexity in CSE is due to code that chooses the best form for
an insn given costs and valid operands for a machine. That's not going
to be affected by tree-ssa. I think it'll need nearly all of the current
hash table mechanism to do that. The "around branches" and related stuff
*should* indeed certainly disappear, but we thought (at least I did) that
it could go away when gcse was written and that turned out not to be the
case. So I'm a bit more skeptical that this will be any different. In any
event, though that code takes up a significant fraction of the compile
*time*, it's not a particularly *complex* part of CSE.